Support expand of types

[campoy]

Since v1.1.0 Dgraph support types and expand(_all_) is defined as the expansion of all the predicates found in all of the types linked to any of the nodes in the query.

So for instance, if we have two types Person and Aged:

type Person {
   first_name: string
   last_name: string
}

type Aged {
   age: int
}

And we have the following dataset:

{
  set {
    _:a <dgraph.type> "Person" .
    _:a <dgraph.type> "Aged" .
    _:a <first_name> "John" .
    _:a <last_name> "Doe" .
    _:a <age> "42" .
     }
}

When we query all values of type Person and use expand(_all_):

{
  q(func: type(Person)) {
    dgraph.type
    expand(_all_)
  }
}

we end up receiving the field age even though it's not part of the type Person because expand(_all_) expands all of the predicates in all of the types linked to any of the values returned.

{
  "data": {
    "q": [
      {
        "dgraph.type": [
          "Person",
          "Aged"
        ],
        "first_name": "John",
        "age": 42,
        "last_name": "Doe"
      }
    ]
  },
  "extensions": {...}
}

This is because our node John Doe is of type Person and Aged, as the result of the query shows.

I propose a way to expand only the fields related to a type by passing the type name as the parameter to expand.

So, we could send this query:

{
  q(func: type(Person)) {
    dgraph.type
    expand(Person)
  }
}

This would return only the fields first_name and last_name. We could also pass multiple types to the expand query: expand(Person, Aged) and that would expand the union of the predicates.

[martinmr]

@campoy How is this different from #3904 It looks like similar functionality but with different syntax.

[martinmr]

Fixed by #3920